254 Professor Arthur Smithells on 



the flame of hydrogen in chlorine and of chlorine in hydrogen. 

 So long as any oxygen remains in the gases the flame shows 

 a distinct yellow colour when a bead of a sodium salt is 

 introduced into it. As the oxygen is removed by a current 

 of purified gas (chlorine or hydrogen), the yellow colour fades 

 until it is imperceptible to the eye. Viewed with the spec- 

 troscope the D line shows a corresponding diminution of 

 intensity, and, in entire absence of oxygen, is doubtless ex- 

 tinguished. As I have only once succeeded in entirely ob- 

 literating it, I do not wish to speak positively. The fact that 

 the red line of lithium is easily quenched under similar con- 

 ditions confirms the belief that the same would be the case 

 with the more sensitive sodium, if all traces of oxygen could 

 be removed. 



It appears to me that the above facts present great 

 chemical anomalies and difficulties. If we are to make 

 the common assumption that the yellow flame is due to 

 sodium vapour, the element must be liberated either by disso- 

 ciation by heat or reduction. Of dissociation by heat we 

 have no independent evidence, and it is gratuitous to suppose 

 that it should take place in a coal-gas flame and not in the 

 flame of hydrogen burning in chlorine. If the process is 

 one of reduction, we have to go beyond our experimental 

 knowledge to suggest an explanation. Thus the liberation 

 of sodium from common salt in a hydrogen-flame must be 

 due either to the union of hydrogen with the chlorine to 

 form hydrochloric acid and free sodium, or to the action of 

 steam which forms hydrochloric acid and sodium hydrate or 

 oxide, the latter being then reduced to the metal by the 

 hydrogen. If it is said that hydrogen may decompose salt in 

 the first way at high temperature, it must also be remarked that 

 hydrochloric acid itself dissociates at high temperatures. If, 

 on the other hand, the steam is the active agent, it is cer- 

 tainly taking a liberty to invoke its presence to act on the 

 salt, and to ignore its probable interference with any reduc- 

 tion of the oxide or hydrate to be subsequently effected by 

 hydrogen. 



If we can explain the coloration of a hydrogen-flame by 

 salt, it will of course be easy to explain it in the case of 

 a coal-gas flame. 



The suppression of the colour in a hydrogen-chlorine flame 

 can only be explained from the point of view of reduction by 

 assuming that the conjoint action of oxygen and hydrogen is 

 necessary for the decomposition of the sodium chloride, the 

 hydrogen taking the chlorine and the oxygen the sodium, the 

 sodium oxide being thereupon reduced by hydrogen. 



